Why We Should Focus on Understanding the Individual in Evaluation

Abstract

This article1 is drawn from the work of Barry MacDonald and is written as a tribute to him. It addresses an issue he raised some while ago whose relevance intensifies today in a world seeing the demise of independent professional practice and public sector ethos. The ‘portrayal of persons’ relates to a paper MacDonald wrote in 1976 (MacDonald 1976), at a time when the discipline and the practice of evaluation was in rapid emergence and development. He was arguing forcefully that to understand innovation we have to understand innovators. This was not a new message. It had been enunciated elsewhere, including by Schwab (1978) and Schön (1971) whose work revealed individualism, values and exchange as key elements of innovation. But what MacDonald added was a political dimension: using the portrayal of persons to share control of the evaluation by passing responsibilities for interpretation (generalisation) to the evaluand; and by empowering the organisation by enhancing knowledge of itself (for example, against attempts to enforce external values on professional action through external accountability). The rationale was the same as that for case study evaluation: to ground generalisations in the experience of the practitioners and to shape evaluation around issues and dilemmas encountered in their practices. This inevitably leads to working with small samples (cases) but not losing the aim of deriving broad generalisation. This was the key methodological principle that provoked and continues to provoke reaction from those who believe in a quantity theory of generalisation—that systems are made up of patterns of action and phenomena, and that we need sufficiently broad ranges of data to ensure that we have captured the full extension of a pattern. Case theorists argue that even where this is true, it is more important to capture the inherent qualities of the patterning and its contingent relationships with context. MacDonald sought to show, as I do in this article, that it is possible to arrive at important and generalisable (provisional) insights from a sample of one.